All terms

Rate limit

A rate limit caps how many API requests one key may make per minute; here it is 120, bursting to 20 requests per second.

In detail

The current state is exposed in the X-RateLimit-* headers on every metered response, so read them instead of counting requests yourself. Exceeding the limit returns 429 with a retry_after value in the body and in the headers; sleep for that long rather than retrying immediately, because the bucket refills continuously. Nearly every limit problem is a polling loop, and long polling or webhooks remove it entirely.

Also known as

  • throttling
  • requests per minute

Related terms